[libc-commits] [libc] [libc] Add max length argument to decimal to float (PR #84091)

[Nick Desaulniers via libc-commits]

================
@@ -296,45 +298,52 @@ class HighPrecisionDecimal {
   }
 
 public:
-  // numString is assumed to be a string of numeric characters. It doesn't
+  // num_string is assumed to be a string of numeric characters. It doesn't
   // handle leading spaces.
-  HighPrecisionDecimal(const char *__restrict numString) {
+  LIBC_INLINE
+  HighPrecisionDecimal(
+      const char *__restrict num_string,
----------------
nickdesaulniers wrote:

Orthogonal to this PR, but putting restrict on a parameter when there's only one doesn't make sense.  restrict should only be used when there's 2 or more parameters of the same type.  Was this addition based on studying disassembly?

https://github.com/llvm/llvm-project/pull/84091